file.c 19.5 KB
Newer Older
C
Chris Mason 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

C
Chris Mason 已提交
19 20 21 22 23 24 25 26 27 28 29 30 31 32
#include <linux/buffer_head.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/smp_lock.h>
#include <linux/backing-dev.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/statfs.h>
#include <linux/compat.h>
33
#include <linux/version.h>
C
Chris Mason 已提交
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "ioctl.h"
#include "print-tree.h"


static int btrfs_copy_from_user(loff_t pos, int num_pages, int write_bytes,
				struct page **prepared_pages,
				const char __user * buf)
{
	long page_fault = 0;
	int i;
	int offset = pos & (PAGE_CACHE_SIZE - 1);

	for (i = 0; i < num_pages && write_bytes > 0; i++, offset = 0) {
		size_t count = min_t(size_t,
				     PAGE_CACHE_SIZE - offset, write_bytes);
		struct page *page = prepared_pages[i];
		fault_in_pages_readable(buf, count);

		/* Copy data from userspace to the current page */
		kmap(page);
		page_fault = __copy_from_user(page_address(page) + offset,
					      buf, count);
		/* Flush processor's dcache for this page */
		flush_dcache_page(page);
		kunmap(page);
		buf += count;
		write_bytes -= count;

		if (page_fault)
			break;
	}
	return page_fault ? -EFAULT : 0;
}

static void btrfs_drop_pages(struct page **pages, size_t num_pages)
{
	size_t i;
	for (i = 0; i < num_pages; i++) {
		if (!pages[i])
			break;
		unlock_page(pages[i]);
		mark_page_accessed(pages[i]);
		page_cache_release(pages[i]);
	}
}

84 85
static int insert_inline_extent(struct btrfs_trans_handle *trans,
				struct btrfs_root *root, struct inode *inode,
86
				u64 offset, ssize_t size,
87
				struct page *page, size_t page_offset)
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
{
	struct btrfs_key key;
	struct btrfs_path *path;
	char *ptr, *kaddr;
	struct btrfs_file_extent_item *ei;
	u32 datasize;
	int err = 0;
	int ret;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	btrfs_set_trans_block_group(trans, inode);

	key.objectid = inode->i_ino;
	key.offset = offset;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
	BUG_ON(size >= PAGE_CACHE_SIZE);
	datasize = btrfs_file_extent_calc_inline_size(size);

	ret = btrfs_insert_empty_item(trans, root, path, &key,
				      datasize);
	if (ret) {
		err = ret;
		goto fail;
	}
	ei = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
	       path->slots[0], struct btrfs_file_extent_item);
	btrfs_set_file_extent_generation(ei, trans->transid);
	btrfs_set_file_extent_type(ei,
				   BTRFS_FILE_EXTENT_INLINE);
	ptr = btrfs_file_extent_inline_start(ei);

123
	kaddr = kmap_atomic(page, KM_USER0);
124
	btrfs_memcpy(root, path->nodes[0]->b_data,
125
		     ptr, kaddr + page_offset, size);
126
	kunmap_atomic(kaddr, KM_USER0);
C
Chris Mason 已提交
127
	btrfs_mark_buffer_dirty(path->nodes[0]);
128 129 130 131 132
fail:
	btrfs_free_path(path);
	return err;
}

C
Chris Mason 已提交
133 134 135 136 137 138 139 140 141
static int dirty_and_release_pages(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct file *file,
				   struct page **pages,
				   size_t num_pages,
				   loff_t pos,
				   size_t write_bytes)
{
	int err = 0;
142
	int i;
C
Chris Mason 已提交
143
	struct inode *inode = file->f_path.dentry->d_inode;
144 145 146 147 148 149 150 151
	struct extent_map *em;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	u64 hint_block;
	u64 num_blocks;
	u64 start_pos;
	u64 end_of_last_block;
	u64 end_pos = pos + write_bytes;
	loff_t isize = i_size_read(inode);
C
Chris Mason 已提交
152

153 154 155
	em = alloc_extent_map(GFP_NOFS);
	if (!em)
		return -ENOMEM;
C
Chris Mason 已提交
156

157
	em->bdev = inode->i_sb->s_bdev;
C
Chris Mason 已提交
158

159 160 161
	start_pos = pos & ~((u64)root->blocksize - 1);
	num_blocks = (write_bytes + pos - start_pos + root->blocksize - 1) >>
			inode->i_blkbits;
C
Chris Mason 已提交
162

163
	down_read(&BTRFS_I(inode)->root->snap_sem);
164
	end_of_last_block = start_pos + (num_blocks << inode->i_blkbits) - 1;
165
	lock_extent(em_tree, start_pos, end_of_last_block, GFP_NOFS);
166 167 168 169 170 171 172 173 174 175 176
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	if (!trans) {
		err = -ENOMEM;
		goto out_unlock;
	}
	btrfs_set_trans_block_group(trans, inode);
	inode->i_blocks += num_blocks << 3;
	hint_block = 0;

	if ((end_of_last_block & 4095) == 0) {
177
		printk("strange end of last %Lu %zu %Lu\n", start_pos, write_bytes, end_of_last_block);
178 179 180 181 182 183 184 185 186 187 188 189
	}
	set_extent_uptodate(em_tree, start_pos, end_of_last_block, GFP_NOFS);

	/* FIXME...EIEIO, ENOSPC and more */

	/* insert any holes we need to create */
	if (inode->i_size < start_pos) {
		u64 last_pos_in_file;
		u64 hole_size;
		u64 mask = root->blocksize - 1;
		last_pos_in_file = (isize + mask) & ~mask;
		hole_size = (start_pos - last_pos_in_file + mask) & ~mask;
C
Chris Mason 已提交
190

191
		if (last_pos_in_file < start_pos) {
C
Chris Mason 已提交
192 193 194 195 196 197 198 199
			err = btrfs_drop_extents(trans, root, inode,
						 last_pos_in_file,
						 last_pos_in_file + hole_size,
						 &hint_block);
			if (err)
				goto failed;

			hole_size >>= inode->i_blkbits;
200 201 202 203 204 205
			err = btrfs_insert_file_extent(trans, root,
						       inode->i_ino,
						       last_pos_in_file,
						       0, 0, hole_size);
		}
		if (err)
C
Chris Mason 已提交
206
			goto failed;
207 208 209 210 211 212 213 214
	}

	/*
	 * either allocate an extent for the new bytes or setup the key
	 * to show we are doing inline data in the extent
	 */
	if (isize >= PAGE_CACHE_SIZE || pos + write_bytes < inode->i_size ||
	    pos + write_bytes - start_pos > BTRFS_MAX_INLINE_DATA_SIZE(root)) {
215
		u64 last_end;
216 217 218
		for (i = 0; i < num_pages; i++) {
			struct page *p = pages[i];
			SetPageUptodate(p);
219
			set_page_dirty(p);
C
Chris Mason 已提交
220
		}
221 222 223 224
		last_end = pages[num_pages -1]->index << PAGE_CACHE_SHIFT;
		last_end += PAGE_CACHE_SIZE - 1;
		set_extent_delalloc(em_tree, start_pos, end_of_last_block,
				 GFP_NOFS);
225 226
	} else {
		struct page *p = pages[0];
227 228
		/* step one, delete the existing extents in this range */
		/* FIXME blocksize != pagesize */
C
Chris Mason 已提交
229 230 231 232 233
		err = btrfs_drop_extents(trans, root, inode, start_pos,
			 (pos + write_bytes + root->blocksize -1) &
			 ~((u64)root->blocksize - 1), &hint_block);
		if (err)
			goto failed;
234

235 236 237 238
		err = insert_inline_extent(trans, root, inode, start_pos,
					   end_pos - start_pos, p, 0);
		BUG_ON(err);
		em->start = start_pos;
239
		em->end = end_pos - 1;
240 241 242 243 244 245 246
		em->block_start = EXTENT_MAP_INLINE;
		em->block_end = EXTENT_MAP_INLINE;
		add_extent_mapping(em_tree, em);
	}
	if (end_pos > isize) {
		i_size_write(inode, end_pos);
		btrfs_update_inode(trans, root, inode);
C
Chris Mason 已提交
247 248
	}
failed:
249 250 251
	err = btrfs_end_transaction(trans, root);
out_unlock:
	mutex_unlock(&root->fs_info->fs_mutex);
252
	unlock_extent(em_tree, start_pos, end_of_last_block, GFP_NOFS);
253
	free_extent_map(em);
254
	up_read(&BTRFS_I(inode)->root->snap_sem);
C
Chris Mason 已提交
255 256 257
	return err;
}

258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end)
{
	struct extent_map *em;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;

	while(1) {
		em = lookup_extent_mapping(em_tree, start, end);
		if (!em)
			break;
		remove_extent_mapping(em_tree, em);
		/* once for us */
		free_extent_map(em);
		/* once for the tree*/
		free_extent_map(em);
	}
	return 0;
}

C
Chris Mason 已提交
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
/*
 * this is very complex, but the basic idea is to drop all extents
 * in the range start - end.  hint_block is filled in with a block number
 * that would be a good hint to the block allocator for this file.
 *
 * If an extent intersects the range but is not entirely inside the range
 * it is either truncated or split.  Anything entirely inside the range
 * is deleted from the tree.
 */
int btrfs_drop_extents(struct btrfs_trans_handle *trans,
		       struct btrfs_root *root, struct inode *inode,
		       u64 start, u64 end, u64 *hint_block)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_leaf *leaf;
	int slot;
	struct btrfs_file_extent_item *extent;
	u64 extent_end = 0;
	int keep;
	struct btrfs_file_extent_item old;
	struct btrfs_path *path;
	u64 search_start = start;
	int bookend;
	int found_type;
	int found_extent;
	int found_inline;
C
Chris Mason 已提交
303
	int recow;
C
Chris Mason 已提交
304

305 306
	btrfs_drop_extent_cache(inode, start, end - 1);

C
Chris Mason 已提交
307 308 309 310
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;
	while(1) {
C
Chris Mason 已提交
311
		recow = 0;
C
Chris Mason 已提交
312 313 314 315 316 317 318 319 320 321 322 323
		btrfs_release_path(root, path);
		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
					       search_start, -1);
		if (ret < 0)
			goto out;
		if (ret > 0) {
			if (path->slots[0] == 0) {
				ret = 0;
				goto out;
			}
			path->slots[0]--;
		}
324
next_slot:
C
Chris Mason 已提交
325 326 327 328 329 330 331
		keep = 0;
		bookend = 0;
		found_extent = 0;
		found_inline = 0;
		extent = NULL;
		leaf = btrfs_buffer_leaf(path->nodes[0]);
		slot = path->slots[0];
332
		ret = 0;
C
Chris Mason 已提交
333 334 335 336
		btrfs_disk_key_to_cpu(&key, &leaf->items[slot].key);
		if (key.offset >= end || key.objectid != inode->i_ino) {
			goto out;
		}
337
		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY) {
C
Chris Mason 已提交
338 339
			goto out;
		}
C
Chris Mason 已提交
340 341 342 343
		if (recow) {
			search_start = key.offset;
			continue;
		}
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
			extent = btrfs_item_ptr(leaf, slot,
						struct btrfs_file_extent_item);
			found_type = btrfs_file_extent_type(extent);
			if (found_type == BTRFS_FILE_EXTENT_REG) {
				extent_end = key.offset +
					(btrfs_file_extent_num_blocks(extent) <<
					 inode->i_blkbits);
				found_extent = 1;
			} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
				found_inline = 1;
				extent_end = key.offset +
				     btrfs_file_extent_inline_len(leaf->items +
								  slot);
			}
		} else {
			extent_end = search_start;
C
Chris Mason 已提交
361 362 363
		}

		/* we found nothing we can drop */
364 365 366 367 368 369 370 371 372 373
		if ((!found_extent && !found_inline) ||
		    search_start >= extent_end) {
			int nextret;
			u32 nritems;
			nritems = btrfs_header_nritems(
					btrfs_buffer_header(path->nodes[0]));
			if (slot >= nritems - 1) {
				nextret = btrfs_next_leaf(root, path);
				if (nextret)
					goto out;
C
Chris Mason 已提交
374
				recow = 1;
375 376 377 378
			} else {
				path->slots[0]++;
			}
			goto next_slot;
C
Chris Mason 已提交
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
		}

		/* FIXME, there's only one inline extent allowed right now */
		if (found_inline) {
			u64 mask = root->blocksize - 1;
			search_start = (extent_end + mask) & ~mask;
		} else
			search_start = extent_end;

		if (end < extent_end && end >= key.offset) {
			if (found_extent) {
				u64 disk_blocknr =
					btrfs_file_extent_disk_blocknr(extent);
				u64 disk_num_blocks =
				      btrfs_file_extent_disk_num_blocks(extent);
				memcpy(&old, extent, sizeof(old));
				if (disk_blocknr != 0) {
					ret = btrfs_inc_extent_ref(trans, root,
					         disk_blocknr, disk_num_blocks);
					BUG_ON(ret);
				}
			}
			WARN_ON(found_inline);
			bookend = 1;
		}
		/* truncate existing extent */
		if (start > key.offset) {
			u64 new_num;
			u64 old_num;
			keep = 1;
			WARN_ON(start & (root->blocksize - 1));
			if (found_extent) {
				new_num = (start - key.offset) >>
					inode->i_blkbits;
				old_num = btrfs_file_extent_num_blocks(extent);
				*hint_block =
					btrfs_file_extent_disk_blocknr(extent);
				if (btrfs_file_extent_disk_blocknr(extent)) {
					inode->i_blocks -=
						(old_num - new_num) << 3;
				}
				btrfs_set_file_extent_num_blocks(extent,
								 new_num);
C
Chris Mason 已提交
422
				btrfs_mark_buffer_dirty(path->nodes[0]);
C
Chris Mason 已提交
423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442
			} else {
				WARN_ON(1);
			}
		}
		/* delete the entire extent */
		if (!keep) {
			u64 disk_blocknr = 0;
			u64 disk_num_blocks = 0;
			u64 extent_num_blocks = 0;
			if (found_extent) {
				disk_blocknr =
				      btrfs_file_extent_disk_blocknr(extent);
				disk_num_blocks =
				      btrfs_file_extent_disk_num_blocks(extent);
				extent_num_blocks =
				      btrfs_file_extent_num_blocks(extent);
				*hint_block =
					btrfs_file_extent_disk_blocknr(extent);
			}
			ret = btrfs_del_item(trans, root, path);
443
			/* TODO update progress marker and return */
C
Chris Mason 已提交
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471
			BUG_ON(ret);
			btrfs_release_path(root, path);
			extent = NULL;
			if (found_extent && disk_blocknr != 0) {
				inode->i_blocks -= extent_num_blocks << 3;
				ret = btrfs_free_extent(trans, root,
							disk_blocknr,
							disk_num_blocks, 0);
			}

			BUG_ON(ret);
			if (!bookend && search_start >= end) {
				ret = 0;
				goto out;
			}
			if (!bookend)
				continue;
		}
		/* create bookend, splitting the extent in two */
		if (bookend && found_extent) {
			struct btrfs_key ins;
			ins.objectid = inode->i_ino;
			ins.offset = end;
			ins.flags = 0;
			btrfs_set_key_type(&ins, BTRFS_EXTENT_DATA_KEY);
			btrfs_release_path(root, path);
			ret = btrfs_insert_empty_item(trans, root, path, &ins,
						      sizeof(*extent));
472 473 474

			if (ret) {
				btrfs_print_leaf(root, btrfs_buffer_leaf(path->nodes[0]));
C
Chris Mason 已提交
475
				printk("got %d on inserting %Lu %u %Lu start %Lu end %Lu found %Lu %Lu keep was %d\n", ret , ins.objectid, ins.flags, ins.offset, start, end, key.offset, extent_end, keep);
476
			}
C
Chris Mason 已提交
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
			BUG_ON(ret);
			extent = btrfs_item_ptr(
				    btrfs_buffer_leaf(path->nodes[0]),
				    path->slots[0],
				    struct btrfs_file_extent_item);
			btrfs_set_file_extent_disk_blocknr(extent,
				    btrfs_file_extent_disk_blocknr(&old));
			btrfs_set_file_extent_disk_num_blocks(extent,
				    btrfs_file_extent_disk_num_blocks(&old));

			btrfs_set_file_extent_offset(extent,
				    btrfs_file_extent_offset(&old) +
				    ((end - key.offset) >> inode->i_blkbits));
			WARN_ON(btrfs_file_extent_num_blocks(&old) <
				(extent_end - end) >> inode->i_blkbits);
			btrfs_set_file_extent_num_blocks(extent,
				    (extent_end - end) >> inode->i_blkbits);

			btrfs_set_file_extent_type(extent,
						   BTRFS_FILE_EXTENT_REG);
			btrfs_set_file_extent_generation(extent,
				    btrfs_file_extent_generation(&old));
			btrfs_mark_buffer_dirty(path->nodes[0]);
			if (btrfs_file_extent_disk_blocknr(&old) != 0) {
				inode->i_blocks +=
				      btrfs_file_extent_num_blocks(extent) << 3;
			}
			ret = 0;
			goto out;
		}
	}
out:
	btrfs_free_path(path);
	return ret;
}

/*
 * this gets pages into the page cache and locks them down
 */
static int prepare_pages(struct btrfs_root *root,
			 struct file *file,
			 struct page **pages,
			 size_t num_pages,
			 loff_t pos,
			 unsigned long first_index,
			 unsigned long last_index,
523
			 size_t write_bytes)
C
Chris Mason 已提交
524 525 526 527 528
{
	int i;
	unsigned long index = pos >> PAGE_CACHE_SHIFT;
	struct inode *inode = file->f_path.dentry->d_inode;
	int err = 0;
529 530 531
	u64 num_blocks;
	u64 start_pos;

532
	start_pos = pos & ~((u64)root->blocksize - 1);
533 534
	num_blocks = (write_bytes + pos - start_pos + root->blocksize - 1) >>
			inode->i_blkbits;
C
Chris Mason 已提交
535 536 537 538 539 540 541

	memset(pages, 0, num_pages * sizeof(struct page *));

	for (i = 0; i < num_pages; i++) {
		pages[i] = grab_cache_page(inode->i_mapping, index + i);
		if (!pages[i]) {
			err = -ENOMEM;
542
			BUG_ON(1);
C
Chris Mason 已提交
543
		}
C
Chris Mason 已提交
544 545
		cancel_dirty_page(pages[i], PAGE_CACHE_SIZE);
		wait_on_page_writeback(pages[i]);
546 547 548
		if (!PagePrivate(pages[i])) {
			SetPagePrivate(pages[i]);
			set_page_private(pages[i], 1);
549
			WARN_ON(!pages[i]->mapping->a_ops->invalidatepage);
550
			page_cache_get(pages[i]);
C
Chris Mason 已提交
551
		}
552
		WARN_ON(!PageLocked(pages[i]));
C
Chris Mason 已提交
553 554 555 556 557 558 559 560 561 562 563 564 565
	}
	return 0;
}

static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
				size_t count, loff_t *ppos)
{
	loff_t pos;
	size_t num_written = 0;
	int err = 0;
	int ret = 0;
	struct inode *inode = file->f_path.dentry->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
566 567
	struct page **pages = NULL;
	int nrptrs;
C
Chris Mason 已提交
568 569 570
	struct page *pinned[2];
	unsigned long first_index;
	unsigned long last_index;
571 572 573

	nrptrs = min((count + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE,
		     PAGE_CACHE_SIZE / (sizeof(struct page *)));
C
Chris Mason 已提交
574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590
	pinned[0] = NULL;
	pinned[1] = NULL;
	if (file->f_flags & O_DIRECT)
		return -EINVAL;
	pos = *ppos;
	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
	current->backing_dev_info = inode->i_mapping->backing_dev_info;
	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
	if (err)
		goto out;
	if (count == 0)
		goto out;
	err = remove_suid(file->f_path.dentry);
	if (err)
		goto out;
	file_update_time(file);

591
	pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL);
C
Chris Mason 已提交
592 593 594 595 596 597 598 599 600 601 602 603 604

	mutex_lock(&inode->i_mutex);
	first_index = pos >> PAGE_CACHE_SHIFT;
	last_index = (pos + count) >> PAGE_CACHE_SHIFT;

	/*
	 * there are lots of better ways to do this, but this code
	 * makes sure the first and last page in the file range are
	 * up to date and ready for cow
	 */
	if ((pos & (PAGE_CACHE_SIZE - 1))) {
		pinned[0] = grab_cache_page(inode->i_mapping, first_index);
		if (!PageUptodate(pinned[0])) {
C
Chris Mason 已提交
605
			ret = btrfs_readpage(NULL, pinned[0]);
C
Chris Mason 已提交
606 607 608 609 610 611 612 613 614
			BUG_ON(ret);
			wait_on_page_locked(pinned[0]);
		} else {
			unlock_page(pinned[0]);
		}
	}
	if ((pos + count) & (PAGE_CACHE_SIZE - 1)) {
		pinned[1] = grab_cache_page(inode->i_mapping, last_index);
		if (!PageUptodate(pinned[1])) {
C
Chris Mason 已提交
615
			ret = btrfs_readpage(NULL, pinned[1]);
C
Chris Mason 已提交
616 617 618 619 620 621 622 623 624
			BUG_ON(ret);
			wait_on_page_locked(pinned[1]);
		} else {
			unlock_page(pinned[1]);
		}
	}

	while(count > 0) {
		size_t offset = pos & (PAGE_CACHE_SIZE - 1);
625 626
		size_t write_bytes = min(count, nrptrs *
					(size_t)PAGE_CACHE_SIZE -
627
					 offset);
C
Chris Mason 已提交
628 629 630
		size_t num_pages = (write_bytes + PAGE_CACHE_SIZE - 1) >>
					PAGE_CACHE_SHIFT;

631
		WARN_ON(num_pages > nrptrs);
C
Chris Mason 已提交
632 633 634
		memset(pages, 0, sizeof(pages));
		ret = prepare_pages(root, file, pages, num_pages,
				    pos, first_index, last_index,
635
				    write_bytes);
636 637
		if (ret)
			goto out;
C
Chris Mason 已提交
638 639 640

		ret = btrfs_copy_from_user(pos, num_pages,
					   write_bytes, pages, buf);
641 642 643 644
		if (ret) {
			btrfs_drop_pages(pages, num_pages);
			goto out;
		}
C
Chris Mason 已提交
645 646 647 648

		ret = dirty_and_release_pages(NULL, root, file, pages,
					      num_pages, pos, write_bytes);
		btrfs_drop_pages(pages, num_pages);
649 650
		if (ret)
			goto out;
C
Chris Mason 已提交
651 652 653 654 655 656

		buf += write_bytes;
		count -= write_bytes;
		pos += write_bytes;
		num_written += write_bytes;

657
		balance_dirty_pages_ratelimited_nr(inode->i_mapping, num_pages);
C
Chris Mason 已提交
658 659 660 661 662
		btrfs_btree_balance_dirty(root);
		cond_resched();
	}
	mutex_unlock(&inode->i_mutex);
out:
663
	kfree(pages);
C
Chris Mason 已提交
664 665 666 667 668 669 670 671 672 673 674 675 676 677
	if (pinned[0])
		page_cache_release(pinned[0]);
	if (pinned[1])
		page_cache_release(pinned[1]);
	*ppos = pos;
	current->backing_dev_info = NULL;
	return num_written ? num_written : err;
}

static int btrfs_sync_file(struct file *file,
			   struct dentry *dentry, int datasync)
{
	struct inode *inode = dentry->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
678
	int ret = 0;
C
Chris Mason 已提交
679 680 681
	struct btrfs_trans_handle *trans;

	/*
682 683
	 * check the transaction that last modified this inode
	 * and see if its already been committed
C
Chris Mason 已提交
684 685
	 */
	mutex_lock(&root->fs_info->fs_mutex);
686 687 688 689 690 691 692 693 694 695 696 697
	if (!BTRFS_I(inode)->last_trans)
		goto out;
	mutex_lock(&root->fs_info->trans_mutex);
	if (BTRFS_I(inode)->last_trans <=
	    root->fs_info->last_trans_committed) {
		BTRFS_I(inode)->last_trans = 0;
		mutex_unlock(&root->fs_info->trans_mutex);
		goto out;
	}
	mutex_unlock(&root->fs_info->trans_mutex);

	/*
698 699
	 * ok we haven't committed the transaction yet, lets do a commit
	 */
C
Chris Mason 已提交
700 701 702 703 704 705 706
	trans = btrfs_start_transaction(root, 1);
	if (!trans) {
		ret = -ENOMEM;
		goto out;
	}
	ret = btrfs_commit_transaction(trans, root);
out:
707
	mutex_unlock(&root->fs_info->fs_mutex);
C
Chris Mason 已提交
708 709 710
	return ret > 0 ? EIO : ret;
}

C
Chris Mason 已提交
711
static struct vm_operations_struct btrfs_file_vm_ops = {
712 713 714 715 716 717
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23)
	.nopage         = filemap_nopage,
	.populate       = filemap_populate,
#else
	.fault		= filemap_fault,
#endif
C
Chris Mason 已提交
718 719 720 721 722 723 724 725 726 727
	.page_mkwrite	= btrfs_page_mkwrite,
};

static int btrfs_file_mmap(struct file	*filp, struct vm_area_struct *vma)
{
	vma->vm_ops = &btrfs_file_vm_ops;
	file_accessed(filp);
	return 0;
}

C
Chris Mason 已提交
728 729 730
struct file_operations btrfs_file_operations = {
	.llseek		= generic_file_llseek,
	.read		= do_sync_read,
C
Chris Mason 已提交
731
	.aio_read       = generic_file_aio_read,
C
Chris Mason 已提交
732
	.write		= btrfs_file_write,
C
Chris Mason 已提交
733
	.mmap		= btrfs_file_mmap,
C
Chris Mason 已提交
734 735
	.open		= generic_file_open,
	.fsync		= btrfs_sync_file,
C
Christoph Hellwig 已提交
736
	.unlocked_ioctl	= btrfs_ioctl,
C
Chris Mason 已提交
737
#ifdef CONFIG_COMPAT
C
Christoph Hellwig 已提交
738
	.compat_ioctl	= btrfs_ioctl,
C
Chris Mason 已提交
739 740 741
#endif
};